Intelligent Virtual Assistant for Voice-Based Access to Computers and the Internet

As computers and the Internet have continually evolved and permeated our daily lives, they have had an increasing reliance on visually appealing content, thereby becoming less accessible to the visually impaired community. This paper aims to remedy this situation by proposing an intelligent virtual assistant (IVA) that provides voice-based access to desktop and Internet features. Users can converse with the IVA as with another human and can intuitively call for different features as the IVA recognizes phrases that are semantically similar to predefined textual cues. The IVA makes features commonly required in professional and personal settings, such as opening software, surfing the Internet, accessing YouTube, composing emails and file management, available using only voice commands. Some features included in the IVA, found lacking in alternative implementations, are the screen reader and image caption model included that improve web accessibility. Significant accuracy of the features and reliability in task automation has been ensured by comprehensive testing.